xml_col_finder 0.1.0 → 0.2.2
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- checksums.yaml.gz.sig +0 -0
- data/lib/xml_col_finder.rb +175 -31
- data.tar.gz.sig +0 -0
- metadata +2 -2
- metadata.gz.sig +0 -0
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: b98fdd5a28880c8560256ae8a01f5e0a64a3067f2c0974f5c0e668243ce7ad92
|
4
|
+
data.tar.gz: c1177f458579fb3fd7a9d975512d14f5b0e2395ea4cbaa21dc40253b564b78f3
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 73daefd3b3bb3873e21848aad39fd7d9bb570229c8369bb71920bce81474d312b6eb83a9647cf03d66bc4b095788a72dee9925a68d8de8e78554f82155afbdea
|
7
|
+
data.tar.gz: be011c43afe08767e860c436024c65d8b86e507f610ffd7caad367f034cd76dfbfaf5c39882eef9ee6524cc2f90d1de9c564db297f4899df01f3e40e42840885
|
checksums.yaml.gz.sig
CHANGED
Binary file
|
data/lib/xml_col_finder.rb
CHANGED
@@ -24,11 +24,80 @@ class XMLColFinder
|
|
24
24
|
|
25
25
|
end
|
26
26
|
|
27
|
+
#@to_a = a
|
27
28
|
h = group_by_xpath(a)
|
28
29
|
@to_a = truncate_xpath(h).flatten(1)
|
29
30
|
|
30
31
|
end
|
31
32
|
|
33
|
+
def to_code(nametip: true)
|
34
|
+
|
35
|
+
@nametip = nametip
|
36
|
+
@tags = {}
|
37
|
+
|
38
|
+
xpath, remaining = @to_a
|
39
|
+
|
40
|
+
eid = getid(xpath)
|
41
|
+
linex = formatline('doc', eid, xpath)
|
42
|
+
a = scan(remaining, eid)
|
43
|
+
|
44
|
+
lines = a.flatten.compact.prepend linex
|
45
|
+
lines.join("\n").lines\
|
46
|
+
.map {|line| line =~ /.text$/ ? 'puts ' + line : line }.join
|
47
|
+
|
48
|
+
end
|
49
|
+
|
50
|
+
private
|
51
|
+
|
52
|
+
def formatline(pid, eid=nil, key=nil, tail=nil, index: nil)
|
53
|
+
|
54
|
+
if eid then
|
55
|
+
|
56
|
+
nametip = @nametip && tail.is_a?(String)
|
57
|
+
klass = nametip ? key.scan(/@class=['"]([^'"]+)/).last : nil
|
58
|
+
|
59
|
+
line = if klass then
|
60
|
+
desc = klass[0][/^[^\-]+/].gsub(/(?=[A-Z])/,' ').downcase
|
61
|
+
desc += " (e.g. %s)" % [tail.length < 50 ? tail : tail[0..46] + '...']
|
62
|
+
"\n# " + desc + "\n"
|
63
|
+
else
|
64
|
+
''
|
65
|
+
end
|
66
|
+
|
67
|
+
line += "%s = %s.element(\"%s\")" % [eid, pid, key]
|
68
|
+
if tail.is_a? String
|
69
|
+
line += '.text'
|
70
|
+
#line += "\n" if nametip
|
71
|
+
end
|
72
|
+
|
73
|
+
else
|
74
|
+
line = index ? ("%s[%d].text" % [pid, index]) : ("%s.text" % pid)
|
75
|
+
end
|
76
|
+
|
77
|
+
return line
|
78
|
+
end
|
79
|
+
|
80
|
+
|
81
|
+
def getid(rawtag)
|
82
|
+
|
83
|
+
rawtagx = rawtag.split('/').last[/\w+/]
|
84
|
+
|
85
|
+
tag = case rawtagx.to_sym
|
86
|
+
when :a
|
87
|
+
'link'
|
88
|
+
else
|
89
|
+
rawtagx
|
90
|
+
end
|
91
|
+
|
92
|
+
if @tags.include?(tag) then
|
93
|
+
@tags[tag] =~ /\d+$/ ? @tags[tag].succ! : @tags[tag] += '1'
|
94
|
+
else
|
95
|
+
@tags[tag] = tag
|
96
|
+
end
|
97
|
+
|
98
|
+
end
|
99
|
+
|
100
|
+
|
32
101
|
# Groups xpath by matching branches
|
33
102
|
#
|
34
103
|
def group_by_xpath(a)
|
@@ -48,39 +117,39 @@ class XMLColFinder
|
|
48
117
|
|
49
118
|
h2 = {}
|
50
119
|
|
51
|
-
|
52
|
-
while n < a.length
|
120
|
+
a.each do |path,txt|
|
53
121
|
|
54
|
-
path, txt = a[n]
|
55
122
|
stickypath = ''
|
56
123
|
|
57
|
-
|
124
|
+
n = 0
|
125
|
+
while n < path.length
|
58
126
|
|
59
|
-
|
60
|
-
puts "h[+ stickypath + '/' + name]: " \
|
61
|
-
+ h[stickypath + '/' + name].inspect
|
62
|
-
end
|
127
|
+
name = path[n]
|
63
128
|
|
64
|
-
if h[stickypath + '/' + name] > 1
|
129
|
+
if h[stickypath + '/' + name] > 1
|
65
130
|
|
66
131
|
stickypath += '/' + name
|
67
|
-
|
132
|
+
|
133
|
+
if (n == path.length - 1) then
|
134
|
+
|
135
|
+
h2[stickypath.sub(/^\//,'')] ||= []
|
136
|
+
h2[stickypath.sub(/^\//,'')] << txt
|
137
|
+
|
138
|
+
end
|
68
139
|
|
69
140
|
else
|
70
141
|
|
71
|
-
|
72
|
-
h2[stickypath
|
142
|
+
|
143
|
+
h2[stickypath.sub(/^\//,'')] ||= []
|
144
|
+
h2[stickypath.sub(/^\//,'')] << [path.join('/'), txt]
|
73
145
|
break
|
74
146
|
|
75
147
|
end
|
76
|
-
|
148
|
+
n += 1
|
77
149
|
end
|
78
150
|
|
79
|
-
n += 1
|
80
|
-
|
81
151
|
end
|
82
152
|
|
83
|
-
|
84
153
|
if h2.length > 1 then
|
85
154
|
|
86
155
|
a2 = h2.map {|k,v| [k.split('/'), v]}
|
@@ -91,32 +160,107 @@ class XMLColFinder
|
|
91
160
|
return h2
|
92
161
|
|
93
162
|
end
|
163
|
+
|
94
164
|
end
|
95
165
|
|
96
|
-
def
|
166
|
+
def scan(a, eid='doc', pid=eid.clone)
|
167
|
+
|
168
|
+
#puts 'a: ' + a.inspect if @debug
|
169
|
+
|
170
|
+
a.map do |row|
|
171
|
+
|
172
|
+
head, tail = row
|
173
|
+
|
174
|
+
if head.is_a? Array then
|
97
175
|
|
98
|
-
|
176
|
+
hline = scan(row, eid, pid)
|
99
177
|
|
100
|
-
|
101
|
-
len = new_key.length
|
102
|
-
#puts len.inspect
|
103
|
-
puts 'new_key: ' + new_key.inspect if @debug
|
178
|
+
elsif head
|
104
179
|
|
105
|
-
|
180
|
+
if head[0] == '/' then
|
106
181
|
|
107
|
-
|
108
|
-
|
109
|
-
new_len = new_k2.length
|
110
|
-
puts 'new_k2: ' + new_k2.inspect if @debug
|
111
|
-
puts 'v2: ' + v2.inspect if @debug
|
182
|
+
key = head[1..-1]
|
183
|
+
puts 'key: ' + key.inspect if @debug
|
112
184
|
|
113
|
-
|
185
|
+
eid = getid(key)
|
114
186
|
|
115
|
-
|
187
|
+
hline = if tail.is_a? Array and tail.all? {|x| x.is_a? String } then
|
188
|
+
@prev_xpath = true
|
189
|
+
"%s = %s.xpath(\"%s\")" % [eid, pid, key]
|
190
|
+
else
|
191
|
+
@prev_xpath = false
|
192
|
+
formatline(pid, eid, key, tail)
|
193
|
+
end
|
116
194
|
|
195
|
+
else
|
196
|
+
|
197
|
+
hline = formatline(pid=eid)
|
198
|
+
end
|
117
199
|
end
|
118
200
|
|
119
|
-
|
201
|
+
if tail.is_a? Array then
|
202
|
+
|
203
|
+
if tail.compact[0].is_a? Array then
|
204
|
+
|
205
|
+
puts 'tail: ' + tail.inspect if @debug
|
206
|
+
|
207
|
+
tline = scan(tail, eid)
|
208
|
+
|
209
|
+
elsif tail.all? {|x| x.is_a? String} and tail[0][0] != '/'
|
210
|
+
puts '_tail: ' + tail.inspect if @debug
|
211
|
+
tline = tail.map.with_index do |x,i|
|
212
|
+
formatline(pid=eid, index: i)
|
213
|
+
end
|
214
|
+
|
215
|
+
end
|
216
|
+
end
|
217
|
+
|
218
|
+
[hline, tline]
|
219
|
+
end
|
220
|
+
|
221
|
+
end
|
222
|
+
|
223
|
+
|
224
|
+
def truncate_xpath(records, offset=0)
|
225
|
+
|
226
|
+
records.map do |record|
|
227
|
+
|
228
|
+
next unless record
|
229
|
+
if record.is_a? Array and (record[0].is_a? Array or record.length > 2) then
|
230
|
+
|
231
|
+
truncate_xpath(record, offset)
|
232
|
+
|
233
|
+
else
|
234
|
+
|
235
|
+
if record.is_a? String then
|
236
|
+
|
237
|
+
value = record
|
238
|
+
puts 'valuex: ' + value.inspect if @debug
|
239
|
+
value
|
240
|
+
|
241
|
+
else
|
242
|
+
|
243
|
+
key, value = record
|
244
|
+
puts 'key: ' + key.inspect if @debug
|
245
|
+
|
246
|
+
new_key = key[offset..-1]
|
247
|
+
puts 'new_key: ' + new_key.inspect if @debug
|
248
|
+
|
249
|
+
next unless new_key
|
250
|
+
len = new_key.length
|
251
|
+
|
252
|
+
if value.is_a? Array then
|
253
|
+
|
254
|
+
[new_key, truncate_xpath(value, offset+len)]
|
255
|
+
|
256
|
+
else
|
257
|
+
|
258
|
+
puts 'value: ' + value.inspect if @debug
|
259
|
+
[new_key, value]
|
260
|
+
|
261
|
+
end
|
262
|
+
end
|
263
|
+
end
|
120
264
|
|
121
265
|
end
|
122
266
|
|
data.tar.gz.sig
CHANGED
Binary file
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: xml_col_finder
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- James Robertson
|
@@ -35,7 +35,7 @@ cert_chain:
|
|
35
35
|
GSlZ9ilAfm8srTjbZ2cWQyNGGxH+zHQ3Z02c4ZEtgPv/wHjptd1VeBm0P1aemsRA
|
36
36
|
ShsxXxzmzIrRENmpBp3tyR3k
|
37
37
|
-----END CERTIFICATE-----
|
38
|
-
date: 2022-01-
|
38
|
+
date: 2022-01-24 00:00:00.000000000 Z
|
39
39
|
dependencies:
|
40
40
|
- !ruby/object:Gem::Dependency
|
41
41
|
name: rexle
|
metadata.gz.sig
CHANGED
Binary file
|